Currently, the available options appear to be limited to GPT-4.1 mini and GPT-4o mini. While these models are suitable for simple tasks, their accuracy and response quality can be insufficient for more complex customer service, technical, and knowledge base use cases.
Please consider adding support for more capable models, such as the full GPT-4.1 model and newer GPT-5-class models. Ideally, administrators could choose which model to use based on the required quality, speed, and cost.
Access to more advanced models would significantly improve the reliability and usefulness of the AI features.